From 066f6a06305c715c94054ea00734e9259d5a2257 Mon Sep 17 00:00:00 2001 From: Philipp Hagemeister Date: Thu, 23 Jan 2014 00:12:47 +0100 Subject: [PATCH] [nowness] Add support --- youtube_dl/extractor/__init__.py | 1 + youtube_dl/extractor/brightcove.py | 2 +- youtube_dl/extractor/nowness.py | 49 ++++++++++++++++++++++++++++++ 3 files changed, 51 insertions(+), 1 deletion(-) create mode 100644 youtube_dl/extractor/nowness.py diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py index 4d6aeabdf..118982ff8 100644 --- a/youtube_dl/extractor/__init__.py +++ b/youtube_dl/extractor/__init__.py @@ -136,6 +136,7 @@ from .nhl import NHLIE, NHLVideocenterIE from .niconico import NiconicoIE from .ninegag import NineGagIE from .novamov import NovamovIE +from .nowness import NownessIE from .nowvideo import NowVideoIE from .ooyala import OoyalaIE from .orf import ORFIE diff --git a/youtube_dl/extractor/brightcove.py b/youtube_dl/extractor/brightcove.py index e1c45d1f0..443294e6f 100644 --- a/youtube_dl/extractor/brightcove.py +++ b/youtube_dl/extractor/brightcove.py @@ -198,7 +198,7 @@ class BrightcoveIE(InfoExtractor): def _extract_video_info(self, video_info): info = { 'id': compat_str(video_info['id']), - 'title': video_info['displayName'], + 'title': video_info['displayName'].strip(), 'description': video_info.get('shortDescription'), 'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'), 'uploader': video_info.get('publisherName'), diff --git a/youtube_dl/extractor/nowness.py b/youtube_dl/extractor/nowness.py new file mode 100644 index 000000000..b1bcb7e54 --- /dev/null +++ b/youtube_dl/extractor/nowness.py @@ -0,0 +1,49 @@ +from __future__ import unicode_literals + +import re + +from .brightcove import BrightcoveIE +from .common import InfoExtractor +from ..utils import ( + ExtractorError, +) + + +class NownessIE(InfoExtractor): + _VALID_URL = r'https?://(?:www\.)?nowness\.com/[^?#]*?/(?P[0-9]+)/(?P[^/]+?)(?:$|[?#])' + + _TEST = { + 'url': 'http://www.nowness.com/day/2013/6/27/3131/candor--the-art-of-gesticulation', + 'file': '2520295746001.mp4', + 'md5': '0ece2f70a7bd252c7b00f3070182d418', + 'info_dict': { + 'description': 'Candor: The Art of Gesticulation', + 'uploader': 'Nowness', + 'title': 'Candor: The Art of Gesticulation', + } + } + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + video_id = mobj.group('slug') + + webpage = self._download_webpage(url, video_id) + player_url = self._search_regex( + r'"([^"]+/content/issue-[0-9.]+.js)"', webpage, 'player URL') + real_id = self._search_regex( + r'\sdata-videoId="([0-9]+)"', webpage, 'internal video ID') + + player_code = self._download_webpage( + player_url, video_id, + note='Downloading player JavaScript', + errnote='Player download failed') + player_code = player_code.replace("'+d+'", real_id) + + bc_url = BrightcoveIE._extract_brightcove_url(player_code) + if bc_url is None: + raise ExtractorError('Could not find player definition') + return { + '_type': 'url', + 'url': bc_url, + 'ie_key': 'Brightcove', + }